PyPI - natural-pdf - Versions diffs - 0.1.7__py3-none-any.whl → 0.1.9__py3-none-any.whl - Mend

natural-pdf 0.1.7py3-none-any.whl → 0.1.9py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Files changed (134) hide show

natural_pdf/__init__.py +3 -0
natural_pdf/analyzers/layout/base.py +1 -5
natural_pdf/analyzers/layout/gemini.py +61 -51
natural_pdf/analyzers/layout/layout_analyzer.py +40 -11
natural_pdf/analyzers/layout/layout_manager.py +26 -84
natural_pdf/analyzers/layout/layout_options.py +7 -0
natural_pdf/analyzers/layout/pdfplumber_table_finder.py +142 -0
natural_pdf/analyzers/layout/surya.py +46 -123
natural_pdf/analyzers/layout/tatr.py +51 -4
natural_pdf/analyzers/text_structure.py +3 -5
natural_pdf/analyzers/utils.py +3 -3
natural_pdf/classification/manager.py +422 -0
natural_pdf/classification/mixin.py +163 -0
natural_pdf/classification/results.py +80 -0
natural_pdf/collections/mixins.py +111 -0
natural_pdf/collections/pdf_collection.py +434 -15
natural_pdf/core/element_manager.py +83 -0
natural_pdf/core/highlighting_service.py +13 -22
natural_pdf/core/page.py +578 -93
natural_pdf/core/pdf.py +912 -460
natural_pdf/elements/base.py +134 -40
natural_pdf/elements/collections.py +712 -109
natural_pdf/elements/region.py +722 -69
natural_pdf/elements/text.py +4 -1
natural_pdf/export/mixin.py +137 -0
natural_pdf/exporters/base.py +3 -3
natural_pdf/exporters/paddleocr.py +5 -4
natural_pdf/extraction/manager.py +135 -0
natural_pdf/extraction/mixin.py +279 -0
natural_pdf/extraction/result.py +23 -0
natural_pdf/ocr/__init__.py +5 -5
natural_pdf/ocr/engine_doctr.py +346 -0
natural_pdf/ocr/engine_easyocr.py +6 -3
natural_pdf/ocr/ocr_factory.py +24 -4
natural_pdf/ocr/ocr_manager.py +122 -26
natural_pdf/ocr/ocr_options.py +94 -11
natural_pdf/ocr/utils.py +19 -6
natural_pdf/qa/document_qa.py +0 -4
natural_pdf/search/__init__.py +20 -34
natural_pdf/search/haystack_search_service.py +309 -265
natural_pdf/search/haystack_utils.py +99 -75
natural_pdf/search/search_service_protocol.py +11 -12
natural_pdf/selectors/parser.py +431 -230
natural_pdf/utils/debug.py +3 -3
natural_pdf/utils/identifiers.py +1 -1
natural_pdf/utils/locks.py +8 -0
natural_pdf/utils/packaging.py +8 -6
natural_pdf/utils/text_extraction.py +60 -1
natural_pdf/utils/tqdm_utils.py +51 -0
natural_pdf/utils/visualization.py +18 -0
natural_pdf/widgets/viewer.py +4 -25
{natural_pdf-0.1.7.dist-info → natural_pdf-0.1.9.dist-info}/METADATA +17 -3
natural_pdf-0.1.9.dist-info/RECORD +80 -0
{natural_pdf-0.1.7.dist-info → natural_pdf-0.1.9.dist-info}/WHEEL +1 -1
{natural_pdf-0.1.7.dist-info → natural_pdf-0.1.9.dist-info}/top_level.txt +0 -2
docs/api/index.md +0 -386
docs/assets/favicon.png +0 -3
docs/assets/favicon.svg +0 -3
docs/assets/javascripts/custom.js +0 -17
docs/assets/logo.svg +0 -3
docs/assets/sample-screen.png +0 -0
docs/assets/social-preview.png +0 -17
docs/assets/social-preview.svg +0 -17
docs/assets/stylesheets/custom.css +0 -65
docs/document-qa/index.ipynb +0 -435
docs/document-qa/index.md +0 -79
docs/element-selection/index.ipynb +0 -915
docs/element-selection/index.md +0 -229
docs/finetuning/index.md +0 -176
docs/index.md +0 -170
docs/installation/index.md +0 -69
docs/interactive-widget/index.ipynb +0 -962
docs/interactive-widget/index.md +0 -12
docs/layout-analysis/index.ipynb +0 -818
docs/layout-analysis/index.md +0 -185
docs/ocr/index.md +0 -209
docs/pdf-navigation/index.ipynb +0 -314
docs/pdf-navigation/index.md +0 -97
docs/regions/index.ipynb +0 -816
docs/regions/index.md +0 -294
docs/tables/index.ipynb +0 -658
docs/tables/index.md +0 -144
docs/text-analysis/index.ipynb +0 -370
docs/text-analysis/index.md +0 -105
docs/text-extraction/index.ipynb +0 -1478
docs/text-extraction/index.md +0 -292
docs/tutorials/01-loading-and-extraction.ipynb +0 -194
docs/tutorials/01-loading-and-extraction.md +0 -95
docs/tutorials/02-finding-elements.ipynb +0 -340
docs/tutorials/02-finding-elements.md +0 -149
docs/tutorials/03-extracting-blocks.ipynb +0 -147
docs/tutorials/03-extracting-blocks.md +0 -48
docs/tutorials/04-table-extraction.ipynb +0 -114
docs/tutorials/04-table-extraction.md +0 -50
docs/tutorials/05-excluding-content.ipynb +0 -270
docs/tutorials/05-excluding-content.md +0 -109
docs/tutorials/06-document-qa.ipynb +0 -332
docs/tutorials/06-document-qa.md +0 -91
docs/tutorials/07-layout-analysis.ipynb +0 -288
docs/tutorials/07-layout-analysis.md +0 -66
docs/tutorials/07-working-with-regions.ipynb +0 -413
docs/tutorials/07-working-with-regions.md +0 -151
docs/tutorials/08-spatial-navigation.ipynb +0 -508
docs/tutorials/08-spatial-navigation.md +0 -190
docs/tutorials/09-section-extraction.ipynb +0 -2434
docs/tutorials/09-section-extraction.md +0 -256
docs/tutorials/10-form-field-extraction.ipynb +0 -512
docs/tutorials/10-form-field-extraction.md +0 -201
docs/tutorials/11-enhanced-table-processing.ipynb +0 -54
docs/tutorials/11-enhanced-table-processing.md +0 -9
docs/tutorials/12-ocr-integration.ipynb +0 -604
docs/tutorials/12-ocr-integration.md +0 -175
docs/tutorials/13-semantic-search.ipynb +0 -1328
docs/tutorials/13-semantic-search.md +0 -77
docs/visual-debugging/index.ipynb +0 -2970
docs/visual-debugging/index.md +0 -157
docs/visual-debugging/region.png +0 -0
natural_pdf/templates/finetune/fine_tune_paddleocr.md +0 -415
natural_pdf/templates/spa/css/style.css +0 -334
natural_pdf/templates/spa/index.html +0 -31
natural_pdf/templates/spa/js/app.js +0 -472
natural_pdf/templates/spa/words.txt +0 -235976
natural_pdf/widgets/frontend/viewer.js +0 -88
natural_pdf-0.1.7.dist-info/RECORD +0 -145
notebooks/Examples.ipynb +0 -1293
pdfs/.gitkeep +0 -0
pdfs/01-practice.pdf +0 -543
pdfs/0500000US42001.pdf +0 -0
pdfs/0500000US42007.pdf +0 -0
pdfs/2014 Statistics.pdf +0 -0
pdfs/2019 Statistics.pdf +0 -0
pdfs/Atlanta_Public_Schools_GA_sample.pdf +0 -0
pdfs/needs-ocr.pdf +0 -0
{natural_pdf-0.1.7.dist-info → natural_pdf-0.1.9.dist-info}/licenses/LICENSE +0 -0

pdfs/.gitkeep DELETED Viewed

File without changes

pdfs/01-practice.pdf DELETED Viewed

@@ -1,543 +0,0 @@
-%PDF-1.7
-%µ¶
-1 0 obj
-<</Type/Pages/Count 1/Kids[2 0 R]>>
-endobj
-2 0 obj
-<</Type/Page/MediaBox[0 0 612 792]/Resources 3 0 R/Parent 1 0 R/Contents 6 0 R>>
-endobj
-3 0 obj
-<</Font<</F0<</Type/Font/BaseFont/Helvetica/Subtype/Type1>>/F1<</Type/Font/BaseFont/Helvetica-Bold/Subtype/Type1>>>>>>
-endobj
-4 0 obj
-<</Length 4482>>
-stream
-q
-    1 0 0 1 50 700 cm
-    1 g
-    325 65 180 -35 re
-    S
-Q
-q
-    1 0 0 1 50 700 cm
-    BT
-        335 50 Td
-        /F0 8 Tf
-        (Jungle Health and Safety Inspection Service) Tj
-        0 -10 Td
-        1 0 0 rg
-        (INS-UP70N51NCL41R) Tj
-    ET
-Q
-q
-    1 0 0 1 50 700 cm
-    BT
-        /F1 10 Tf
-        (Site: ) Tj
-        /F0 10 Tf
-        (Durham's Meatpacking  ) Tj
-        0.5 g
-        (Chicago, Ill.) Tj
-        0 g
-        0 -20 Td
-        /F1 10 Tf
-        (Date:  ) Tj
-        /F0 10 Tf
-        (February 3, 1905) Tj
-        0 -20 Td
-        /F1 10 Tf
-        (Violation Count: ) Tj
-        /F0 10 Tf
-        (7) Tj
-        0 -20 Td
-        /F1 10 Tf
-        (Summary: ) Tj
-        % 0 -20 Td
-        /F0 10 Tf
-        (Worst of any, however, were the fertilizer men, and those who served in the cooking rooms.) Tj
-        0 -16 Td
-        (These people could not be shown to the visitor - for the odor of a fertilizer man would scare any ordinary ) Tj
-        0 -16 Td
-        (visitor at a hundred yards, and as for the other men, who worked in tank rooms full of steam, and in ) Tj
-        0 -16 Td
-        (some of which there were open vats near the level of the floor, their peculiar trouble was that they fell) Tj
-        0 -16 Td
-        (into the vats; and when they were fished out, there was never enough of them left to be worth ) Tj
-        0 -16 Td
-        (exhibiting - sometimes they would be overlooked for days, till all but the bones of them had gone out) Tj
-        0 -16 Td
-        (to the world as Durham's Pure Leaf Lard!) Tj
-    ET
-Q
-q
-    1 0 0 1 50 440 cm
-    0 G
-    2 w
-    0 0 m
-    500 0 l
-    S
-    BT
-    0 -30 Td
-    /F1 12 Tf
-    (Violations) Tj
-    ET
-Q
-q
-    1 0 0 1 50 400 cm
-    0.5 G
-    0 0 m
-    500 0 l
-    S
-    0 -20 m
-    500 -20 l
-    S
-    0 -40 m
-    500 -40 l
-    S
-    0 -60 m
-    500 -60 l
-    S
-    0 -80 m
-    500 -80 l
-    S
-    0 -100 m
-    500 -100 l
-    S
-    0 -120 m
-    500 -120 l
-    S
-    0 -140 m
-    500 -140 l
-    S
-    0 -160 m
-    500 -160 l
-    S
-    0 0 m
-    0 -160 l
-    S
-    50 0 m
-    50 -160 l
-    S
-    400 0 m
-    400 -160 l
-    S
-    450 0 m
-    450 -160 l
-    S
-    500 0 m
-    500 -160 l
-    S
-Q
-q
-    1 0 0 1 55 386 cm
-    BT
-        q
-            /F1 10 Tf
-            (Statute) Tj
-            50 0 Td
-            (Description) Tj
-            350 0 Td
-            (Level) Tj
-            50 0 Td
-            (Repeat?) Tj
-        Q
-        /F0 10 Tf
-        q
-            0 -20 Td
-            (4.12.7) Tj
-            50 0 Td
-            (Unsanitary Working Conditions.) Tj
-            350 0 Td
-            (Critical) Tj
-        Q
-        q
-            0 -40 Td
-            (5.8.3) Tj
-            50 0 Td
-            (Inadequate Protective Equipment.) Tj
-            350 0 Td
-            (Serious) Tj
-        Q
-        q
-            0 -60 Td
-            (6.3.9) Tj
-            50 0 Td
-            (Ineffective Injury Prevention.) Tj
-            350 0 Td
-            (Serious) Tj
-        Q
-        q
-            0 -80 Td
-            (7.1.5) Tj
-            50 0 Td
-            (Failure to Properly Store Hazardous Materials.) Tj
-            350 0 Td
-            (Critical) Tj
-        Q
-        q
-            0 -100 Td
-            (8.9.2) Tj
-            50 0 Td
-            (Lack of Adequate Fire Safety Measures.) Tj
-            350 0 Td
-            (Serious) Tj
-        Q
-        q
-            0 -120 Td
-            (9.6.4) Tj
-            50 0 Td
-            (Inadequate Ventilation Systems.) Tj
-            350 0 Td
-            (Serious) Tj
-        Q
-        q
-            0 -140 Td
-            (10.2.7) Tj
-            50 0 Td
-            (Insufficient Employee Training for Safe Work Practices.) Tj
-            350 0 Td
-            (Serious) Tj
-        Q
-    ET
-    q
-        0 G
-        0.5 w
-        465 -20 8 8 re
-        S
-        465 -20 m
-        473 -12 l
-        s
-        465 -12 m
-        473 -20 l
-        s
-        465 -40 8 8 re
-        S
-        465 -40 m
-        473 -32 l
-        s
-        465 -32 m
-        473 -40 l
-        s
-        465 -60 8 8 re
-        S
-        465 -80 8 8 re
-        S
-        465 -100 8 8 re
-        S
-        465 -120 8 8 re
-        S
-        465 -120 m
-        473 -112 l
-        s
-        465 -112 m
-        473 -120 l
-        s
-        465 -140 8 8 re
-        S
-    Q
-Q
-q
-    1 0 0 1 230 20 cm
-    BT
-        /F0 8 Tf
-        (Jungle Health and Safety Inspection Service) Tj
-    ET
-Q
-endstream
-endobj
-5 0 obj
-<</Type/Catalog/Pages 1 0 R>>
-endobj
-6 0 obj
-<</Length 2774>>
-stream
-q
-1 0 0 1 50 700 cm
-1 g
-325 65 180 -35 re
-S
-Q
-q
-1 0 0 1 50 700 cm
-BT
-335 50 Td
-/F0 8 Tf
-(Jungle Health and Safety Inspection Service) Tj
-0 -10 Td
-1 0 0 rg
-(INS-UP70N51NCL41R) Tj
-ET
-Q
-q
-1 0 0 1 50 700 cm
-BT
-/F1 10 Tf
-(Site: ) Tj
-/F0 10 Tf
-(Durham's Meatpacking  ) Tj
-.5 g
-(Chicago, Ill.) Tj
-0 g
-0 -20 Td
-/F1 10 Tf
-(Date:  ) Tj
-/F0 10 Tf
-(February 3, 1905) Tj
-0 -20 Td
-/F1 10 Tf
-(Violation Count: ) Tj
-/F0 10 Tf
-(7) Tj
-0 -20 Td
-/F1 10 Tf
-(Summary: ) Tj
-/F0 10 Tf
-(Worst of any, however, were the fertilizer men, and those who served in the cooking rooms.) Tj
-0 -16 Td
-(These people could not be shown to the visitor - for the odor of a fertilizer man would scare any ordinary ) Tj
-0 -16 Td
-(visitor at a hundred yards, and as for the other men, who worked in tank rooms full of steam, and in ) Tj
-0 -16 Td
-(some of which there were open vats near the level of the floor, their peculiar trouble was that they fell) Tj
-0 -16 Td
-(into the vats; and when they were fished out, there was never enough of them left to be worth ) Tj
-0 -16 Td
-(exhibiting - sometimes they would be overlooked for days, till all but the bones of them had gone out) Tj
-0 -16 Td
-(to the world as Durham's Pure Leaf Lard!) Tj
-ET
-Q
-q
-1 0 0 1 50 440 cm
-0 G
-2 w
-0 0 m
-500 0 l
-S
-BT
-0 -30 Td
-/F1 12 Tf
-(Violations) Tj
-ET
-Q
-q
-1 0 0 1 50 400 cm
-.5 G
-0 0 m
-500 0 l
-S
-0 -20 m
-500 -20 l
-S
-0 -40 m
-500 -40 l
-S
-0 -60 m
-500 -60 l
-S
-0 -80 m
-500 -80 l
-S
-0 -100 m
-500 -100 l
-S
-0 -120 m
-500 -120 l
-S
-0 -140 m
-500 -140 l
-S
-0 -160 m
-500 -160 l
-S
-0 0 m
-0 -160 l
-S
-50 0 m
-50 -160 l
-S
-400 0 m
-400 -160 l
-S
-450 0 m
-450 -160 l
-S
-500 0 m
-500 -160 l
-S
-Q
-q
-1 0 0 1 55 386 cm
-BT
-q
-/F1 10 Tf
-(Statute) Tj
-50 0 Td
-(Description) Tj
-350 0 Td
-(Level) Tj
-50 0 Td
-(Repeat?) Tj
-Q
-/F0 10 Tf
-q
-0 -20 Td
-(4.12.7) Tj
-50 0 Td
-(Unsanitary Working Conditions.) Tj
-350 0 Td
-(Critical) Tj
-Q
-q
-0 -40 Td
-(5.8.3) Tj
-50 0 Td
-(Inadequate Protective Equipment.) Tj
-350 0 Td
-(Serious) Tj
-Q
-q
-0 -60 Td
-(6.3.9) Tj
-50 0 Td
-(Ineffective Injury Prevention.) Tj
-350 0 Td
-(Serious) Tj
-Q
-q
-0 -80 Td
-(7.1.5) Tj
-50 0 Td
-(Failure to Properly Store Hazardous Materials.) Tj
-350 0 Td
-(Critical) Tj
-Q
-q
-0 -100 Td
-(8.9.2) Tj
-50 0 Td
-(Lack of Adequate Fire Safety Measures.) Tj
-350 0 Td
-(Serious) Tj
-Q
-q
-0 -120 Td
-(9.6.4) Tj
-50 0 Td
-(Inadequate Ventilation Systems.) Tj
-350 0 Td
-(Serious) Tj
-Q
-q
-0 -140 Td
-(10.2.7) Tj
-50 0 Td
-(Insufficient Employee Training for Safe Work Practices.) Tj
-350 0 Td
-(Serious) Tj
-Q
-ET
-q
-0 G
-.5 w
-465 -20 8 8 re
-S
-465 -20 m
-473 -12 l
-s
-465 -12 m
-473 -20 l
-s
-465 -40 8 8 re
-S
-465 -40 m
-473 -32 l
-s
-465 -32 m
-473 -40 l
-s
-465 -60 8 8 re
-S
-465 -80 8 8 re
-S
-465 -100 8 8 re
-S
-465 -120 8 8 re
-S
-465 -120 m
-473 -112 l
-s
-465 -112 m
-473 -120 l
-s
-465 -140 8 8 re
-S
-Q
-Q
-q
-1 0 0 1 230 20 cm
-BT
-/F0 8 Tf
-(Jungle Health and Safety Inspection Service) Tj
-ET
-Q
-endstream
-endobj
-xref
-0 7
-0000000000 65536 f
-0000000016 00000 n
-0000000068 00000 n
-0000000165 00000 n
-0000000300 00000 n
-0000004833 00000 n
-0000004879 00000 n
-trailer
-<</Size 7/Root 5 0 R>>
-startxref
-7704
-%%EOF

pdfs/0500000US42001.pdf DELETED Viewed

Binary file

pdfs/0500000US42007.pdf DELETED Viewed

Binary file

pdfs/2014 Statistics.pdf DELETED Viewed

Binary file

pdfs/2019 Statistics.pdf DELETED Viewed

Binary file

pdfs/Atlanta_Public_Schools_GA_sample.pdf DELETED Viewed

Binary file

pdfs/needs-ocr.pdf DELETED Viewed

Binary file

{natural_pdf-0.1.7.dist-info → natural_pdf-0.1.9.dist-info}/licenses/LICENSE RENAMED Viewed

File without changes

natural-pdf 0.1.7__py3-none-any.whl → 0.1.9__py3-none-any.whl

natural-pdf 0.1.7py3-none-any.whl → 0.1.9py3-none-any.whl